Impacts of New Health Information on Health Behaviours

In our research we attempt to understand how people change their behaviours in response to their partners receiving new health information. The COVID-19 pandemic has brought the importance of health information into the limelight, but do people actually respond to new health information? Although some previous work has found that health information does play a role in determining behaviours, in our research we don’t find any evidence of that. Although we do find evidence that partners change behaviours, we find that there is an alternative causal mechanism that is the main driving factor in these behavioural changes.

Based on previous work we know that there is a very high correlation between partners in terms of health, both mental and physical, as well as in terms of behaviours. Indeed, previous work has found high correlations in lifestyle health-related behaviours like physical activity, diet, tobacco consumption, and alcohol consumption. This might be expected, but there have been three theories proposed to explain this phenomenon: assortative matching, shared environment, and joint household production.

The basic idea behind assortative matching is that people choose to match with people similar to them in terms of their likes and dislikes. An economist would frame this as preferences being complements. In other words, you would choose to “match” with someone precisely because you have shared likes or hobbies. This is intuitive, because people may meet their partner at their tennis club or go on their first few dates to a cinema because they both enjoy films. Having similar likes or interests does explain some of the correlation in partners’ behaviours, but not all of it.

The second explanation is shared environment. The idea here is that those living in the same household will behave similarly because the environment (broadly defined) determines behaviours. Although they may not respond in the same way to their environment, they still live in the same place as one another. In the form of an example, we could think about running in the rain. Although my partner and I hate running in the rain, my partner is more tolerant to the rain, consequently, will run more often when it rains. However, we will do a similar amount of running as we are both exposed to the same number of rainy days because we live in the same place. An important component of this shared environment theory is shared information. It seems reasonable that couples would share information. Although they might not respond to this information in the same way, they still have access to similar information. If someone is told something by their physician about a particular condition, it would be reasonable to think they would share this information with their partner, so that they both have access to similar information. Sharing this information would then explain some of the correlation in their behaviours.

Finally, joint household production is the theory that a change in the way one household member behaves will change how other household members behave. This theory is probably the most economic heavy one, however thinking about this as joint participation is a more intuitive way of understanding this channel. In the form of an example, let’s consider a typical household that decides to cook one meal for all household members because it is more efficient to do so. If one household member decides to eat more vegetables, then we would expect all household members to eat more vegetables because of the efficiency of cooking one shared meal. An alternative example would be that partners like spending time together. If one household member decides to start running or walking in the evening, then, they their partner may join them for their evening run, because they enjoy spending time together, even if they hate running. The key point here is that partners behave in a similar way because it is efficient to do so.

These theories provide relatively intuitive reasons why we might find partners behaving in a very similar way. These theories provide an opportunity to investigate how partners might change their behaviour if one was to receive new health information. The latter two of these theories suggest that new health information given to one household member might induce changes in the behaviours of other household members, and this is what we investigate. To be specific, we analyse what impacts a diabetes diagnosis has on the partner of the diagnosed person. Theoretically, if someone was diagnosed with diabetes, then their physician would likely tell them about the disease, and they would then share this information with their partner. This may induce a change in their partner’s behaviour because they are now more aware of the consequences of a diabetes diagnosis. Alternatively, we might think that someone diagnosed with diabetes decides to eat less sugary meals, this would also change their partner’s behaviour because, as discussed above, it might be more efficient to cook a healthy meal to share, rather than cook separately.

In our analysis, we find that partners of those that receive a diabetes diagnosis do change their behaviour. We find physical activity increases and they decrease their tobacco consumption. We think this, in itself, is interesting, but what may be more interesting is understanding which of the two theoretical channels this effect works through.

When analysing these channels, we find that joint household production (or more intuitively joint participation) explains the entire effect, and information transfer plays no role in determining these behaviours. What this means, in practice, is that new information about diabetes does not explain the impacts on partner’s behaviour. Now we may, or may not, be concerned about this. On one hand, maybe we shouldn’t be too worried about this. People may already be aware of the risks of diabetes, and they already choose to partake in behaviours based on their well-informed position. However, if people are not responsive to new information, and physicians or policymakers want to reduce the prevalence of diabetes, what other tools do they have at their disposal? Of course, our research focuses on quite a specific context, and more work needs to be done in this area to fully understand how people respond. Is the information on some conditions more impactful than others? Or are there specific types of information that elicit more meaningful responses?